USING LANDSAT THEMATIC MAPPER IMAGERYTO SUPPORT SALMON RESTORATION EFFORTS

IN A LARGE PACIFIC COAST WATERSHED

Jan Derksen, Ph.D., HeadInformation Systems Development

William M. Kier Associates

Lawrence Fox Ill, Ph.D., ChairForestry Department

Humboldt State University

Ronald A. Iverson, Ph.D., Project LeaderKlamath River Fish and Wildlife Office

U.S. Fish and Wildlife Service

June, 1996

William M. Kier AssociatesSausalito, California

CONTENTS

Introduction

Information is key to Klamath salmon restoration cooperationThe Klamath Landsat vegetative data users list grows

Met hods

Mapping the basins 1994 vegetative conditionDetecting changes in vegetation cover over a 20-year period

Results 6

GIS, map analysisDetecting changes in vegetation cover over a 20-year period

References

2

3

35

915

19

Table 1. The habitat type classification system used to map existing vegetationfrom 1994 Landsat imagery.

Figure 1. 1994 vegetation map of the Scott River sub-basin.

Figure 2. 1994 vegetation map of the Shasta River sub-basin.

Figure 3. Relative percentage of land area by vegetation type, Scott and Shasta sub-basins

Figure 4. 1994 vegetation map of the Scott River sub-basin constrained to a 90-m. buffer

Figure 5. 1994 vegetation map of the Shasta River sub-basin constrained to a 90-m. buffer

Figure 6. Area summaries of vegetation classesthe Scott and Shasta river sub-basins

within a 90-m. buffer around streams in

Figure 7. Spectral change, 1973 to 1984, South Fork Mountain, northwestern California

Figure 8. Spectral change, 1984 to 1992, South Fork Mountain, northwestern California

USING LANDSAT THEMATIC MAPPER IMAGERYTO SUPPORT SALMON RESTORATION EFFORTS

IN A LARGE PACIFIC COAST WATERSHED

Jan Derksen, Ph.D, Lawrence Fox 111, Ph.D3 and Ronald Iverson, Ph.D

INTRODUCTION

Pacific salmon populations have declined precipitously in the past three decades. Theirdecline has devastated salmon-dependent communities from central California to PugetSound. Congress has authorized several initiatives to restore salmon populations,including the 1986 Klamath Act (Public Law 99-552). The Act directs the Secretary of theInterior to carry out a 20-year-long Klamath River Basin Fishery Restoration Program. Toassist the Secretary in carrying out the program the Act also established a Klamath RiverFisheries Task Force of federal, state and local government officials, commercial, anglingand tribal fishery stakeholders.

The Klamath Task Force prepared a Long Range Plan for the Klamath Riser BasinConservation Area Fishery Restoration Program. The plan organized the availableinformation concerning the Klamaths fish, fisheries, and fish habitat and identified thefactors responsible for the basins salmon population decline. Chief among these appearedto be a long-term decline in water quality from non-point pollution sources. Loggingand grazing appear to have reduced streamside vegetation which, in turn, increased streamtemperatures, accelerated bank erosion and smothered streambeds with sediments.

Information Is Key to Klamath Salmon Restoration Cooperation

While the plans findings were clear to the Klamaths fish-interested agencies and parties,local landowners and politicians naturally resisted suggestions that traditional watersheduses had contributed to the salmon decline and must, therefore, be modified if the salmonare to be restored. These local interests could easily veto the plan by questioning theimpacts of grazing and other land and water uses and denying access to the priority streamrestoration sites. It was a situation that would require, among other things, abundantinformation about the condition of the streams and their fish populations. Moreover, theinformation would have to be presented to the communities in a clear and unequivocalmanner.

Project enabled by Interagency Agreement W-18-567 between the National Aeronautics and Space

2Administration and the U.S. Fish and Wildlife Service, Region 1. Portland.

3William M. Kier Associates. 2015 Bridgeway, Suite 304. Sausalito. CA 94965Forestry Department. Humboldt State University, Arcata. CA 95521

4 U.S. Fish and Wildlife Service. P.O. Box 1006. Yreka. CA 96097

Anticipating the continuing need for information to sustain the Restoration Programsmonitoring, evaluation and adaptive management needs, as well as the need to convincethe broader community of the necessity of specific restoration measures, the Task Forcesplan called for the development of a basinwide coordinated information system. Further,the plan specifically recommended that the Task Force evaluate the U.S. EnvironmentalProtection Agencys Stream Reach file as the information systems hydrographic base andLandsat data as a tool for long term change detection, both back in time and forwardthrough the Restoration Programs life.

The coordinated information system was begun in 1992 and dubbed the Klamath ResourceInformation System, or KRIS. Reach File data (version 3, or RF3) was acquired soonthereafter and is being refined to give each stream segment in the huge (10 million acre)watershed, as well as each piece of information about that reach, its own digital address.Finally, Landsat Thematic Mapper (TM) imagery became available to the RestorationProgram in 1994 through Interagency Agreement W- 18-567 between the NationalAeronautics and Space Administration and the U.S. Fish and Wildlife Service. The Serviceadministers the Restoration Program on behalf of the Secretary and the Task Force.

The Interagency Agreement proposes that the TM data be used to classify the Klamathbasins existing vegetative condition, the specific conditions in the stream corridors thathave been prioritized for restoration, and the changes that have occurred in thosecorridors since Landsat data first became available in the early 1970s.

The Klamath Landsat Vegetative Data Users List Grows

While the specific intent of our Landsat project was to create an information element thatwould strengthen KRIS power to support stream and fish population restoration work, ithas become increasingly clear that our 1994 Klamath basin vegetative data layer will bevaluable to a number of ecosystem restoration efforts now underway in the region. Thiscomes as no surprise since vegetation reflects so many of a sites physical factors,including climate, soil type, elevation, aspect and recent disturbance. It reflects anecosystems primary producers and the kinds and extent of terrestrial habitat. In short, anexisting vegetation condition map is an indispensable tool for ecosystem assessment andrestoration (Scott, et al., 1993).

In addition to the Klamath Fishery Restoration Program the Landsat TM-derivedvegetative layer will support the Fish and Wildlife Services work related to the NorthwestForest Plan, the Department of the Interiors Klamath Basin Ecosystem RestorationInitiative, the U.S. Bureau of Reclamations Klamath Reclamation Project OperationsPlan and the upper Klamath basin ecological restoration projects contemplated in theOregon Resource Conservation Act of 1996, S. 1662. The data has been prepared forintegration, as well, in the State of Californias California Environmental ResourcesEvaluation System (CERES), the National Gap Analysis program, as well as other Stateand federal spatial databases for Oregon and California.

2

METHODS

We processed 1994 Landsat Thematic Mapper imagery by computer to create maps ofexisting vegetation cover throughout selected watersheds. Vegetation maps were overlaidwith stream maps to assess near-stream conditions. Near-stream vegetation cover wassummarized within a 90 meter stream buffer, thus creating a corridor of analysis 90 meterswide on both sides of each stream, a 180-meter wide total corridor. We compared theshade-producing potential of near stream vegetation in two sub-basins within the KlamathRiver hydrobasin. In addition to this 1994 vegetation mapping, we also detected changesin vegetation cover over a 20-year time span (1973 - 1992) using co-registered LandsatMSS imagery from 1973, 1984, and 1992. Changes were detected within a region knownto have experienced disturbance by fire and logging over this time period in order todemonstrate the utility of spectral change detection for quantifying changes in landscapecondition that may cumulatively affect fish habitat.

Mapping the Basins 1994 Vegetative Condition

Landsats Thematic Mapper provided electronic, computer-compatible imagery with aground resolution of 30 meters and a nominal picture element (pixel) size of 30 meters by30 meters (0.09 hectares). The imagery was acquired by the Landsat satellite betweenJune 22, 1994 and August 9, 1994 with five of the images acquired between July 8 and 17,1994. The close proximity of the acquisition dates provided seamless coverage duringone, early-summer season of plant phenology in a time-frame surrounding the summersolstice, thereby minimizing the effects of terrain shadowing in the imagery. We developedseparate spectral signatures for each Landsat scene except for scenes acquired on the sameday, for which signatures were developed across multiple scenes. It was necessary todevelop unique signatures for each day of image acquisition due to the slightly differentspectral properties of the atmosphere on different days.

Our spectral classification system was patterned after Californias Wildlife HabitatRelationships (WHR) Classification System (Mayer and Laudenslayer 1988). We couldnot match the WHR classification system exactly however because spectral signaturesfrom the Landsat Thematic Mapper sensor sometimes failed to discriminate specific WHRhabitat types and some